Merge lp:~dpigott/lava-dispatcher/increase-timeout into lp:lava-dispatcher

[Dave Pigott]

Have only increased the timeout. Didn't want to destabilise while move was going on. Need to open another bug, which will basically allow us to have a timeout parameter on deploy_linaro_image (and deploy_linaro_android_image for that matter)

[Paul Larson (pwlars)]

I'm going to wait on this, not intending to merge as-is for 12.03

[Dave Pigott (dpigott)]

I've added https://bugs.launchpad.net/lava-dispatcher/+bug/961926 for the actual timeout implementation

[Zygmunt Krynicki (zyga)]

+1

review: Approve

[Zygmunt Krynicki (zyga)]

I've merged this to 0.5.12 hot-fix release and released to pypi
I've also merged this to trunk (after bumping trunk to 0.6.dev)
Finally I've deployed this to production

=== modified file 'lava_dispatcher/client/master.py'
--- lava_dispatcher/client/master.py	2012-03-20 06:59:13 +0000
+++ lava_dispatcher/client/master.py	2012-03-21 13:35:36 +0000
@@ -81,7 +81,10 @@
     session.run('udevadm trigger')
     session.run('mkdir -p /mnt/root')
     session.run('mount /dev/disk/by-label/testrootfs /mnt/root')
-    _deploy_tarball_to_board(session, rootfs, '/mnt/root', timeout=3600)
+    # The timeout has to be this long for vexpress. For a full desktop it
+    # takes 214 minutes, plus about 25 minutes for the mkfs ext3, add
+    # another hour to err on the side of caution.
+    _deploy_tarball_to_board(session, rootfs, '/mnt/root', timeout=18000)
 
     session.run('echo linaro > /mnt/root/etc/hostname')
     #DO NOT REMOVE - diverting flash-kernel and linking it to /bin/true